论文部分内容阅读
目的探索新型农村合作医疗的损失分布模型并简要介绍其SAS实现方法。方法选取常用的几种损失分布模型,利用SAS 9.1软件OR模块中的NLP过程分别进行参数估计并利用χ2检验和Kolmogorov-Smirnov法分别进行拟合优度检验。结果镇级、市级和市外定点医疗机构住院次数的最优分布模型为二阶混合Poisson分布,市外非定点医疗机构住院次数的最优分布模型为Poison-二项分布;镇级、市级、市外定点和市外非定点医疗机构每次费用额的最优分布模型分别为Log-Logistic分布、Log-Logistic分布、Burr分布和Burr分布。结论 SAS在新型农村合作医疗的损失分布拟合研究中具有可行性和可靠性。
Objective To explore the loss distribution model of new rural cooperative medical system and briefly introduce its SAS implementation method. Methods Several commonly used loss distribution models were selected, and the parameters were estimated respectively using the NLP process of SAS 9.1 software OR module and the goodness of fit test was performed using χ 2 test and Kolmogorov-Smirnov method respectively. Results The optimal distribution model of inpatient number at fixed-point medical institutions in town, city and city was second-order mixed Poisson distribution. Poisson-binomial distribution was the optimal distribution model of hospitalization number of non-designated medical institutions outside the city. Logistic distribution, Log-Logistic distribution, Burr distribution and Burr distribution are the best distribution models for each cost of fixed-point and non-fixed-point medical institutions outside the city. Conclusion SAS is feasible and reliable in the study of the fitting of loss distribution in new rural cooperative medical system.